Seed Study: Polyphasic Sleep in Ten Steps

by LoganStrohl3 min read11th Jul 2013135 comments


SleepSelf Experimentation

(Update on this project now available here.)

A handful of Bay Area folks will be going polyphasic over the next month. By that, I mean we'll be adopting a sleep schedule that gets us 4 extra hours of productive work or play time per day, or two whole months per year. (Or a decade over 60 years.)

If you want to tell me about why it's a bad idea, feel free to post comments. I don't plan to use this space to sell you on polyphasic sleep. That might be another post, depending on how this goes.

I'm going to be collecting some very simple data through this here form. I invite you to join us!

This will be hard. It will hurt. You'll probably need a buddy to follow you around and keep you awake. If you don't have a lot of self-discipline, I don't recommend even trying.

Still with me? If you want in by the time you're done reading this, message me (or comment below) with your name so I know who you are. Here's the plan.

1.   Stop using caffeine right now. If you try to maintain a caffeine addition during this process, you will fail. I promise.

2.   Data collection began on July 10th. Start submitting daily reports at any point as soon as you want to participate, especially if you can begin in the next couple of days and then stick to our schedule. Fill out the above form once every 24hrs (whenever it's convenient) until August 10th.

3.   Pick a time to take a 20min nap each day from Monday, July 15th through Sunday, July 21st. You probably won't actually sleep during this time, but you can use it for mindfulness meditation if you stay awake. The goal is to practice napping. This is important.

4.   On Monday, July 22nd, begin fasting immediately after lunch.

5.   On the night of Monday, July 22nd, skip sleep. No naps, then an all-nighter. This is the official adaptation start date. The idea is to make you sleep deprived so your naps the next day are more likely to take.

6.   Eat breakfast on the morning of Tuesday, July 23rd. This should be the first time you've eaten anything since Monday lunch.

7.   Starting on the morning of Tuesday, July 23rd, take a 20min nap every 2hrs (for a total of 12 naps per day). Do not oversleep. Use an obnoxious alarm or whatever other means necessary. "Nap" counts as lying down trying to sleep; take your naps on a strict schedule regardless of how long you successfully sleep.

8.   Start to cut your naps down toward 6 a day as quickly as you can without it hurting too much. Beginning to dream during your naps is a good indicator that you're ready for this part.*

9.   Once you're down to one nap every 4 hours, you're on what's known as the Uberman schedule.

10.  Matt Fallshaw informs me that the next part is a little tricky.

10.1.   If you managed to reach the Uberman feeling good, you'll probably start getting really tired again shortly thereafter. This flavor of tired will be different from what you've suffered for the past week, and by that flavor you will know that you have hit SWS deprivation. If this is what happens to you, the new kind of sleepy is your cue to transition straight to the Everyman 3 schedule, which means a 3 hour block of core sleep plus three 20 minute naps spaced evenly throughout the day. And that's it!

10.2.   If you're unlucky, you'll not quite have reached Uberman in the space of a week--that is, you'll still be hanging on to some extra naps on July 30th. Then you'll be wolloped by a new bout of sleepiness. This flavor of tired will be different from the last. If it's is tolerable, drop straight to full Uberman and try to hold out for at least 24hrs, then convert to the Everyman 3. If the new flavor of tired is intolerable, convert to E3 as soon as the new tired hits, and expect the next week or so to be tougher on you than on the lucky ones.

Why are we doing this weird naptation adaptation plan thing instead of just going straight for the Everyman 3? Mostly because Matthew Fallshaw said to. If you know Matt, that's enough. In case you don't: It takes people about a month to adapt to the Everyman 3, but only about a week to adapt to the Uberman. The Uberman forces your body to learn to get its REM and SWS in those tiny 20 minute naps. If you're still giving it core sleep time, your body won't take the fullest possible advantage of naps right away.

If you think you can keep the Uberman schedule indefinitely, go for it! But keep me informed about it so I know what's up with my data.

*A clarification from Matt: "Drop naps as quickly as you can while remaining functional. The most important part of this period is that you don't sleep for longer than 20 minutes at a time, but the earlier you can get to a pure Uberman schedule the better. Take naps as you need them (with at least 40 minutes awake and moving around between naps) while pushing towards Uberman. The longer you can maintain pure Uberman before introducing a longer core sleep block the further along you'll be to a full adaptation."

ETA: I chose the psychomotor vigilance task (which you'll find if you check out the form I linked to) because the specific thing I'm trying to do here is distinguish polyphasic from chronic partial sleep restriction. If people on polyphasic return to their monophasic PVT baseline after a couple of weeks, especially if they stay there for a long time, that's clear evidence that the polyphasers are not experiencing the same physiological phenomenon as people suffering from chronic partial sleep restriction, which is what I'm actually concerned about. One of the only really well established facts in the literature on partial sleep restriction is that people who are deprived of a couple of hours of sleep a night get worse at the PVT as a function of time. If it's the case that the polyphasers will end up with memory problems, attention problems, and related issues, the simplest explanation is that they're suffering from a kind of chronic partial sleep restriction. I hope that clears some things up.


135 comments, sorted by Highlighting new comments since Today at 10:27 AM
New Comment
Some comments are truncated due to high volume. (⌘F to expand all)Change truncation settings

Please do mental benchmarking: Mnemosyne/Anki/spaced-repetition, dual n-back, , - something objective.

I'll be doing several of these, but what everyone's doing is a psychomotor vigilance task, which my research so far suggests is the most efficient way known to track the whole jumble of things under the heading of "sleep deprivation". If the results of this are interesting and I do a much bigger thing in the future, we'll do a whole lot more tests then.

I'll be doing several of these, but what everyone's doing is a psychomotor vigilance task, which my research so far suggests is the most efficient way known to track the whole jumble of things under the heading of "sleep deprivation".

The PVT is a good thing to use. For people not familiar with it, it tracks the ability to maintain focus, which is the primary thing that sleep deprivation destroys (and makes driving while tired so dangerous). There are long-term concerns about mental ability which the PVT will not address- some people who have done these sorts of sleep schedules find that they run out of 'creativity' after a few weeks, which is a serious concern but something hard to measure. (It's also hard for me to differentiate between "I generate ideas for 8 hours of useful work a day, and when I'm awake 20 hours a day that's obviously too little" and "I don't generate ideas anymore" from their reports.)

6LoganStrohl8yRight, there are a whole bunch of concerns, some easier to test than others. I chose the PVT because I wanted a test that distinguishes effects of a polyphasic schedule from effects of chronic partial sleep restriction. And it now occurs to me as blindingly obvious that I should have said that outright in the OP.
2Jonathan_Graehl8yPVT is good for public safety. I'd like to have stats from some of the cambridge games as well. PVT is a sort of bare minimum mental performance measure. Because I sometimes do challenging work, I care sometimes about quality hours awake, not adequate/alert hours.
1ChristianKl8yI don't think dual n-back is a good idea. I can't imagine people under the stress of the adaption to actually stick with it. [] is probably much better.

Prediction: Since you are making this significantly harder than what it needs to be most people will give up before they even do Everyman for a week. And nobody will sustain Uberman since you aren't even committing to trying to keep it up.

Unrelated: Bay Area rationalists seem to rely a lot on arguments from authority over everything else(I don't have any data but I just see comments like the one about Matt relatively often). I wonder if seeing awesome people that are right a lot of the time might be coming with a cost.

5JGWeissman8yThe initial week of uberman (with extra naps) is actually a lot easier that transitioning straight to everyman 3. The point is to learn to get REM sleep with 20 minute naps without interference from a 3 hour core providing REM. By attempting a nap every 2 hours, you make failed attempts at napping less frustrating because there are more opportunities to try, and even a failed attempt can be quite refreshing (but not a sustainable replacement for sleep) if you practice mindfulness meditation. In comparison, waking from a 3 hour core when you are not getting good naps just leaves you wanting to go back to sleep. The initial week of just napping without cores may sound more ambitious than going straight to everyman 3, but it really is an easier adaption.
2Tenoke8yI understood the rationale behind it but the question is do you have any evidence to back this up? Sounds like armchair theorizing to me. For example, when I used to do everyman I would want to go back to sleep after a nap almost as much as after the core (and quite quickly I would actually start feeling somewhat refreshed from the core). However, I wouldn't generalize from my experience too much. Also you are adding a no-sleep (not too much) and fasting on the 22nd but doing 12 naps a day and all kinds of stuff like that - no strict schedule for the transition from the 12 naps a day to 6 naps a day for example.Which is not necessarily counter-productive but it is at least questionable as it is unusual. And I am not even mentioning how you are giving yourself way more options to fail and other such potential drawbacks to your 'study'.

I understood the rationale behind it but the question is do you have any evidence to back this up? Sounds like armchair theorizing to me.

You're completely right; I don't have solid evidence to back this up. In fact, that's why I'm doing this. I have various anecdotes and largely irrelevant sleep studies. This is preliminary exploratory research, to see whether it's worth my time to look into it further. It's whatever the opposite of 'armchair theorizing' is.

4JGWeissman8yI have actually tried this method of adapting and it worked. I don't have a large amount of data, that is why people are organizing to keep track of it. This adaption method is based on the observation that people can adapt straight to uberman faster than they can adapt straight to everyman 3.
0Tenoke8yThen why aren't the people doing this by going straight to Uberman? Also did you do straight to uberman or straight to 12*20mins naps - there is a decent difference between the two
9MalcolmOcean8yThere are several phases to the everyman adaptation. 1: Get tired. The body needs REM sleep a lot. Like, within days. It's like water. This means that once you've spent a few days without REM, the body will try sneak it in during tiny naps. 2a: Learn to REM nap. Reason #1 why to do 12 naps/day is to give yourself twice as many chances to practice. 2b: Get tired... again? During phase 2, you're also depriving yourself of deep sleep. Your body can last a bit longer without deep sleep, which is why this phase takes awhile. However, you want to be getting as much REM sleep as possible, so that you aren't also REM-deprived. 3: Compressed deep sleep. Finally, when you can't take it any longer, you give yourself several hours to sleep in the early night, when the body most wants to deep sleep. The 12-nap method is also recommended for uberman. If you go straight into uberman, then even if you're ultimately getting enough sleep, you probably aren't getting extra, which means that since you started sleep deprived then you're going to stay sleep deprived. Extra naps help overcome this. I did exactly this back in April, and the first few nights of deep sleep (phase 3) were incredible. Here's a photo [] of my Zeo showing that I had nearly a solid block of deep sleep for 2.5 hours. This isn't desirable long-term (cycles are healthy) but it's cool to see. Most people would get more like 1h of deep sleep in the first 3 hours of sleep, and maybe 1.5h total. For more information, see this page about adaptation on Polyphasic Society [].

There are no double-blind studies. All evidence so far is largely anecdotal, because it's extremely challenging to get a statistically significant number of people to do this to themselves at the same time. In addition to being painful, it's also disruptive to everyday life for awhile. There is a myriad of statistical evidence about how sleep deprivation works, but by and large most people just report "if you don't get X hours of sleep per night, then things suck".

Based on many anecdotes (which are spread around in personal correspondence, forums, listservs, etc) the instructions above make sense. It is the best data available and the point of this pilot study is to build momentum towards doing a more in-depth study. But we have to start with a hypothesis, and based on the anecdotes, this is that hypothesis.

I've spent somewhere between 20 and 50 hours reading about polyphasic sleep, and have tried two adaptations, one with each method. My uberman adaptation using the 6nap one (with sometimes 1 extra) failed, but years later I successfully adapted to everyman using the 12nap method. So this collective experience means that I have evidence that this adaptation plan is a solid one, it's just not easy to impart to you.

ETA: My point here is that if we insist on having evidence before we do experiments, we will not do a lot of science today.

0Tenoke8yI didn't ask for double-blind studies. I was saying that I understand the rationale behind it ( so no need to repeat it it again) but am not convinced unless there is some evidence. How is this related? We are talking about this method versus more 'traditional' methods of adapting to everyman. Granted, I have not been heavily involved in the related communities for the last ~5 years but I have seen significantly more people adapting everyman through the standard method of just jumping in (or jumping in and making some slight alterations at least). Fwiw, if you can provide the information that you are talking about here it would've sufficed to some extent as evidence (or at least data). You are assuming that based on the anectodes this is the optimal(ish) hypothesis but you have not provided them. If you wish I can link you (after some googling) to a lot of people who claim to have achieved everyman through more traditional methods for example. The biggest difference between the two attempts is that in one you were going for uberman and in the second you were going for everyman. This, I suspect makes a bigger difference than the use of 6 versus 12 naps. Who exactly is insisting on that??
7ModusPonies8yI'm not sure what this argument is about. I don't think you two actually disagree on any questions of fact.
0MalcolmOcean8yYeah, and I'm basically done with it. The double-blind studies remark was sarcasm, which I was hoping would be apparent given the impracticality of deceiving people about whether they were doing 6 or 12 naps :P
2MalcolmOcean8yI really don't have the energy to respond to all of this, but I'll point out that the main reason that you probably haven't heard of many people doing the exaptation/naptation method in general is that it's a recent (~2y) innovation []. ETA: To elaborate, Bayesically, the question we want to be asking is not "of the successful adapters, how many used which method?" but "of the people who attempt each method, what proportion are successful?" Not that we can hope to get an accurate answer to either by googling. What we do have is the advice of Matt Fallshaw (who has coached a number of people through successful adaptations (myself and JGWeissman included) and he recommends this method. We could ask him how many of his coachees have been successful. And then for more on going uberman=>everyman, see this page on Polyphasic Society []:
0MalcolmOcean8yThis, and the fact that my body probably sort of remembered how to do the REM naps from last time. I accidentally suggested that the success itself was solid evidence. I don't believe it is. What I meant to do was simply cite my own experience with this stuff (in addition to my research) which strongly suggests more naps (well-spaced) would not have any negative physiological effects, and would have a net positive psychological effect (because you really want to sleep, so it's nice to let yourself sleep more often). Therefore a good idea.
1JGWeissman8yI started out at 12 20 minute naps. This allowed more opportunities to practice napping, with less time in between forcing myself to stay awake. Once I started dreaming during naps, I reduced to strict uberman. During the transition, it was a nap every two hours at night, and every 3 or 4 hours during daylight, as needed. I think there are strong arguments for extra naps initially. If we can get a large experimental group, it might make sense to randomly assign some to start with 12 naps and some to start with 6, to get some data about it. If this could be arranged, I would bet that 12 naps group would have an easier adaption.
0Tenoke8yI would bet on the 6 nap group if this actually happens at some point (which I strongly doubt). How long did it take you from the very start of taking up polyphasic to getting used to everyman? Also have you tried the normal method (starting straight with everyman) or 6 naps from the start?
0JGWeissman8yI switched to everyman 3 on the 9th night of my adaption (counting the initial sleepless night as the first night). I have only adapted once, and I don't think it would be useful to do it again, because now I already know how to nap. If an experiment is setup which defines a measurement of success, we should make a bet on it then.
0Tenoke8yI'm sorry but my question was: by this I mean how long did it take you from the first night of no sleep/naps only to being adapted to everyman and not feeling sleep deprived. Unless you are claiming that you were already adapted on the first day after the switch?
0JGWeissman8yFrom my first day actually on everyman 3, I was getting more quality awake hours than on monophasic. I was actually doing really great initially, until I got the flu a week in, and then it took a while to reestablish the schedule. I don't really remember the timeline very well.
2Tenoke8yI am skeptical as to whether you were immediately adjusted as opposed to feeling better because you added a lot of sleep to your prior schedule (uberman). Getting the flu (or just having flu-like symptoms) and falling out of schedule seems like further evidence for the second option
0MalcolmOcean8yIf I were in the study, I would pay to be in the 12 naps group (I realize this would be bad data wise). Sleep dep sucks, and taking extra naps can be such a relief. So I'd bet that for both physiological and psychological reasons, the 12 naps group would be much more likely to succeed, and to enjoy it.
0Puredoxyk7yOf course, 12 naps means twice the opportunities to oversleep... I'm fascinated by the idea of using extra naps during adaptation, since it is a natural response to want them, and the people I've known who took them deliberately didn't find them harmful (unless they overslept, which is a definite risk if one is already exhausted). I never allowed for them myself.

So, I've tried uberman twice, and failed both times. A report for the first time is here, and I thought I had written up the second time but I'm not finding it easily. The second time, one of my friends tried Everyman, and got the impression that it was physiologically similar to only sleeping 5-6 hours a night; he had more time to do stuff, but was less rested for that time.

I strongly recommend having someone else there to make sure you're awake when you should be. Otherwise, you very likely will crash and sleep for eight hours. Many people have the 'answering machine' wake up with their alarm, do whatever is necessary to shut the alarm off, and then go back to sleep, without remembering any of it. This is not an experience I had until I hadn't slept more than 30 minutes in ~five days.

Also, at some point my ability to make coherent sentences took a hit, and I would sometimes say word salad. Since the experiment, I have had occasional word recall problems, which I notice about once a month (though I think the frequency was higher closer to the experiment). Thinking they're connected may be post hoc ergo propter hoc (there are several other plausible explanations), but I'm trying to avoid the failure mode of "well, no one's reported detrimental effects yet, so I should discount my experience."

2[anonymous]8yThis is business as usual for me, but then I have a somewhat abusive relationship with my subconscious. Feeling well-rested and getting 'enough' sleep only occurs when I don't have anything sufficiently interesting to keep me up. That being said, I have tried uberman before, and it was devastating in the short term: I felt very burnt out and constantly fatigued. However it's possible that I just wasn't doing it correctly.
1Vaniver8yThis is the intended result, for some values of "short term." In theory, after approximately a week of only napping, you begin to get REM sleep during naps and then recover. This did not happen to me in 7 days, and I had only arranged to be continuously watched for 7 days, and so I promptly crashed day 8 or 9.

How social are the people trying this? Do any of you have jobs with scheduled anything? Go to classes or events that are 2 hours long? Having "more hours" a day, if works would be great but it seems to be insanely inconvenient for anyone who spends a lot of time driving or wants to watch a movie or go to a concert or needs to work with people. Assuming the final "quality" of the hours you have ends up on average unaffected by sleep deprivation, the extremely weird and disciplined schedule still seems terrible to me, and I don't even have a job that is particularly scheduled. This strikes as something that gives you more lonely hours at the cost of wrecking a lot of the hours you might have with people. This is less of a problem if you're in a community of people doing this but still seems pretty rough.

3James_Miller8ySo there is a market opening for an entire firm or school of people who do this.
2MalcolmOcean8yThat's a huge reason why most people don't stick with uberman. Even uberman, once adapted, allows for 3.5h wake times. Everyman, though, allows one to be quite functional with just a few naps earlier in the day. My current schedule (core 0030-400, naps 0730, 1130, ~330) has the first nap around the time other people are getting up anyway, and then the second during lunch, then the third is fairly flexible, and I have it whenever in the afternoon makes sense. I'm a student; my lunch is the same time every day, but sometimes I have to push back the afternoon nap or skip half a lecture to take it. At any rate, I'm totally free during evening/social hours, aside from not being able to stay out late. It appears [], though, that periodically one can do an earlier core then stay out, or do a late one, and this is not so bad. When I was interning at a tech startup, I would just nap as I needed/wanted. Based on another recent round of interviews, it seems that about a third of tech companies are supportive, a third supportive, and a sixth each reluctant and completely disapproving. Small companies tend to be in the more pro-nap groups, larger companies vary widely. (I didn't keep actual stats on this, so the fractions are rough. The number of companies I asked was <10.)
0ChristianKl8yThat basically why there are no people who stick with the shedule for multiple years.
0aelephant8yThere are ways to adjust your sleep if you miss a nap. I think you typically just extend the core sleep time.

Having done Uberman in the past, I'd like to recommend a few tweaks and general advice.

1) Consider swapping from 20 minute naps to 24 minute naps. The optimal is 24 minutes asleep with 1-2 minutes to fall asleep. I can't seem to find the original article by NASA atm, but here's an article referencing NASA's original study on optimal naps.

2) Devise a program to get from part 7 to part 10 instead of "try to space out longer". I did a standard uberman schedule and going even 20 minutes past my appointed nap time would immediately put me into sleep deprivation which risked oversleeping an alarm. I'm not sure how you plan to soft adjust to longer spaces between naps, but just winging it while sleep deprived seems like a recipe for danger.

3) Stock up on food. Expect to eat 50% more.

4) Make a list of things to do while sleep deprived at 4am for the adjustment period. I thought I was going to learn a new language with Rosetta Stone; I ended up reading fanfiction and watching through TV series. I simply didn't have the mental juice to do anything until after I'd adjusted.

5) Set up a sleep cycle calendar so you'll know what you're going to do every wedge, and what you're doing every day. It's very easy to lose track of days on a polyphasic schedule.

6) Buy a sleep mask. It's so much easier to sleep during the day with one.

3lavalamp7yDisagree about optimal nap length being 24 minutes for everyone. For me, when I was doing polysleeping, anything longer than about 18 minutes of actual sleeping caused my body to switch into a longer sleep cycle. ETA: Use a Zeo to measure this. Or just go by the "time at which you naturally wake up before the 25 minutes expire".
4Puredoxyk7yI disagree as well -- my optimal sleep time is 19 minutes, and I take 1-2 to fall asleep, so I set my alarms for 23 and usually wake up before they go off. I'm not sure this can be formulated in a way that "works for everybody" though. Aiming for 20 minutes of sleep and doing the picky adjustments of a minute here, a minute there after you're adapted seems the most sensible to me.
2marenz7yDo you happen to know an alternative to the Zeo? I googled a bit after I read your comment and the company is out of business since 2012. But there seems to be no alternative device that measures the brain-waves to detect the different phases of sleep.
3lavalamp7yMight be able to get a used one on eBay. Check out: [] and: []
0marenz7yI am not sure I want a device that has no longer any support :/ But the Zzzs thing sounds really cool, I just reserved one for 15$. Thanks for the links!
0diegocaleiro8ySeconding the Sleep mask, or scarf, or Anything that takes the photons away!
0Puredoxyk7yThirding! I LOVE my sleep masks.

I hope none of you polyphasers drive to school/work, or we are likely to be a few rationalists short by the time this experiment is over.

Not necessarily if you weight the number of rationalists by how long each is awake.

Not necessarily if you weight the number of rationalist by how long each is awake.

You also have to weight the quality of their thinking they do while awake, either for the same reason you are multiplying by hours awake or simply because lower quality thinking will push some accross the threshold of qualifying as 'rationalist'.

4Puredoxyk7yI would hope that I'm not the only source that insists on limiting or eliminating driving for at least the few really hard days of an Uberman adaptation, yeah. Also, you know, don't perform surgery or operate giant cranes. Just in case we needed to add that. ;)

Fantastic! Illuminating an evidence-based path to an extra 4 hours of productive work or play time per day would do much to help mankind. Even if you have only, say, a 5% chance of success, the expected value of what you are doing is enormous.

3LoganStrohl8yI agree. Supposing you consistently valued your time at $10/hr, you'd be gaining $280 per week, or $14,560 per year. (Relative to an 8-hour sleep schedule.) And many people I know would add a 0 to the end of those numbers. ETA: Since you mentioned a 5% chance of success, 5% of $10/hr over a year would be $728. (Over sixty years, $43,680.)

I'd like to see study comparing the above approach to regular modafinil use at whatever dose allows alertness with that amount of sleep on an ongoing basis.

Luke Muehlhauser once claimed that, with the help of modafinil, he was able to sleep only once every 48 hours, for a continuous period of a few months. Unfortunately, he never elaborated on how he felt while doing this, or on why he eventually reverted to a normal sleep schedule.

2somervta7yIIRC, he said that after he switched the type of work/things he was doing, it just stopped working/became nonviable.
0Puredoxyk7yOo, so would I! I don't think I'd be willing to be the guinea pig for that though... ;)
[-][anonymous]8y 10

It would be cool if people doing this kept a journal, and wrote down changes in the way they think and feel during this. I tried it twice, and lasted a solid two weeks during one of the attempts, but it wreaked havoc on my ability to store and/or recall memories. I would have a lot of trouble telling you what I did a few days ago, and I would be irrationally angry/upset for about 15 seconds at the person who woke me up.(I did this with a group of three.) A few times I yelled angry nonsense at them for this short time. We had a "safe word" that was supposed to prove we were awake, ours was "alligator". After one nap I was woken up, angrily hissed "ARMADILLO" and tried to go back to sleep. It was pretty funny even at the time, but I decided to stop since I figured it wasn't good for me, and I was pretty miserable. I was definitely micro sleeping at times as well, so I'll echo Petruchio's warning about looking out for that.

Best of luck, hope it works out for you better than it did for me.

2peter_hurford8yI've been keeping a journal of my adventures in polyphasic sleep here []. Right now, I'm not doing anything exciting -- just siesta sleep (6 hour core, 1x20 minute nap), but I intend to eventually shift slowly into Everyman 3, as my schedule allows.

Here are two articles quite skeptical of polyphasic sleep that might be of interest.

I don't necessarily agree with the author, but I guess it's good to be aware of arguments on both sides of the debate.

1Jonathan_Graehl8yI thought this was quite damning. It's been in circulation for a long time (since 2010?) without any refutation. That said, it is clear that there's a tension between (mental) sleep efficiency and sleep debt. Unfortunately, it's easy to demonstrate all kinds of mental impairment from undersleeping.
6Puredoxyk7yActually, I wrote a refutation of it years ago. It's more vitriolic than I'd write if it were today (but then again, today I'd just say that my book contains the refutations for most of this in much more polished form), but it did get answered. Here's the refutation: [] And a follow-up discussion I thought was pretty helpful at the time:!topic/polyphasic/_R4-kdZbpJI [!topic/polyphasic/_R4-kdZbpJI]
0Jonathan_Graehl7yYes, the comments by "a reader" and Michael Turner [!topic/polyphasic/_R4-kdZbpJI] helped me situate Wozniak's (supermemo polyphasic skeptic) point of view. Thanks for the link.
0lavalamp7yIt's very strange to see my past self's words there three years later... (I'm one of the other participants in that thread.) Fortunately I didn't say anything I now regret...

I've a few questions about those "atypical sleep patterns" :

  1. Are there studies about their long-term effect on health, lifespan, IQ, ... ?

  2. How do they cope with sickness/wounds ? If you get the flu, will you be able to heal as fast using a Uberman or Everyman sleep pattern ? People doing monophasic will tend to sleep much more when sick, both increase the size of the monophasic sleep and doing naps. What happens if you follow Uberman/Everyman ? Can you get this additional sleep when sick, without breaking the whole adaptation ?

  3. How do they cop

... (read more)
8MalcolmOcean8yRe: 3... Geoff Anders of Leverage, who has been on Everyman 3 for months, said to me in an email: I think by "polyphasic" he means "adhering to a specific polyphasic schedule" which is what the term often means, given that outside of the polyphasic community people are much more likely to sleep random/flexible hours even if they nap etc.
8LoganStrohl8y1. Nope. Even the literature on chronic partial sleep restriction is abysmal. That's what I'm trying to change. A seed study with a large effect size might get more people with funding taking these questions seriously. 2. Don't know because of 1. 3. See 2. But word on the street is that some people have much more flexibility than others.
2wwa8yI can provide a data point to 3: Not at all! I have 2 days a week where I can't freely schedule my sleep and need to be awake for 8+ hours. This completely killed all attempts to implement any schedule listed on Polyphasic_sleep []. To be fair, I almost made it with Biphasic but I tend to have a "long boot time" so I decided it isn't worth feeling like crap two times a day for half an hour. In the end I invented my own schedule which, again, was only partial success. While experimenting on the schedules I noticed that I sometimes feel very refreshed in the "morning", but not often. I thought it's connected to the REM and tried sleeping multiplies of around 1h30m but either it wasn't connected to REM or my cycle deviates too much from the mean because experimentally I estimated this "good cycle" of mine to be 2h20m. Then on a normal day I'd sleep one "wwa-sleep-cycle" two (for 19.4% sleep time) or three (29.2%) times a day which is not even close to Uberman on average, but I feel awesome. The interesting feature of this schedule of mine is that I can switch in/out of it immediately on the next day. Scaling to external constraints works well as a side effect of adaptation time e.g. before a tough day I'd sleep two cycles continuously or even get back to monophasic sleep with 3 cycles continuously. Of course all this is original research on one specimen so may not apply to anyone else at all. Also, it took me months to do and now I only do it when on a tough deadline because one person would kill me if I'd set an alarm clock in the middle of the night, every night ;-)
0[anonymous]8yHave you tried one of those iPhone/Android apps which guess what part of your sleep cycle you're in based on accelerometer data, and try to wake up when you're not in REM or SWS?
0wwa8yNo, I didn't know it exists, cool. Looks like one more reason to get myself a smartphone already.
0[anonymous]8yThere are widgets that do the same thing [] but are much cheaper than a smartphone¹, though much more expensive than an app if you already have a smartphone. -------------------------------------------------------------------------------- 1. Than a new smartphone, at least -- dunno what the market for second-hand ones is like where you are.
0Fermatastheorem8yHow much did you sleep on a monophasic schedule?
0wwa8yBefore experiments pretty close to 8 on average, maybe slightly less.
8JGWeissman8yWhen I get sick, I add more sleep to my everyman 3 schedule. I can extend naps to 90 minutes or 3 hours, and add extra hours onto my core. In extreme cases, I can just ignore the normal schedule and sleep as much as I feel like. When I notice early that I am getting sick, I can sometimes prevent the worst of it my sleeping for 6 total hours that day. In all cases, I am able to resume normal everyman 3 after recovering.
6Puredoxyk7yThat's been pretty much exactly my experience as well, with the possible addendum that I work really hard to make sure I can sleep as long as I want if I notice that I might be getting sick, since if I catch it early, doing this is VERY likely to prevent the illness altogether.
4Puredoxyk7yStudies, no. I wrote a book (, if ya'll don't mind me dropping that link -- if it's verboten, I'll remove it and sorry) that compiles as much as I've been able to get ahold of as far as information after a decade of running a site and communicating with people on the subject, and it has chapters that address your other two (very good!) questions. The short answer is: AFTER adaptation, polyphasic sleep copes with events (including sickness, travel, and "just life") just like monophasic sleep does, only in a compressed / hyperefficient manner. DURING adaptation it's super strict and will get thrown off by these, but once it's well-ingrained, things work surprisingly similarly -- just shorter.
[-][anonymous]7y 7

The data collection period is now over. How did it go?

5Rob Bensinger7yHere's a preliminary wrap-up: []. A lot more detail to follow.

I have a work schedule and no willing buddy in the area. Could you keep the data form up indefinitely? I'd love to submit to the study at a later date.

0LoganStrohl8yWill do.

I suppose modafinil should be in the same boat as caffeine for the purposes of this experiment.

1Douglas_Knight7yMany people find that modafinil does not interfere with sleep, so it might be compatible with polyphasic sleep, though the threshold of "not interfering" might be different. It is likely to be problematic for this protocol of sleep deprivation.
3RichardKennaway7yIsn't the entire point of modafinil to interfere with sleep?
2LM78057yMy experience with modafinil has been that sleep is possible, but not necessary.
0RichardKennaway7yWhy do it, if it is not necessary?
8LM78057yModafinil eliminates the feeling of being tired, but not the body's need for sleep. Being in sleep deficit weakens the immune system, and I've seen long-term modafinil use without sleep recovery end in pneumonia. So, if I take modafinil to work on a project or stay alert for a long drive, and I finish before the modafinil wears off, I'll go to sleep anyway, because even if my brain doesn't care whether it gets sleep or not, my body needs it. With stimulants like caffeine, getting to sleep before the stimulant wears off is difficult and leads to less sleep.
0Douglas_Knight7ySome people take modafinil to be smarter or concentrate better without changing their sleep schedules. If you were to combine it with polyphasic sleep, the goal would not be to reduce sleep further. Modafinil has a very long half life, 12-16 hours. Stimulants that interfere with sleep require scheduling. The shorter the half-life, the easier the scheduling. This is an advantage of nicotine over caffeine and caffeine over amphetamine. Modafinil has an even longer half life, but it doesn't require knowing your bedtime so far ahead.
0Douglas_Knight7yThe point of modafinil is remove the need for sleep. You might say it interferes with sleepiness. It may be surprising, but it turns out that sleepiness is not necessary for sleep. Indeed, going to bed early when changing time zones is a demonstration of this, but many people find it a challenge. People on modafinil are not sleepy at their normal bedtime, but many of them report not having difficulty getting to sleep if they choose. I think more people on modafinil have problems failing to sleep because they forgot or got distracted than fail after getting in bed. In contrast, most people report that caffeine or amphetamine before bed causes their minds to race and interferes with sleep, making it difficult and low quality.
0Douglas_Knight7yI changed my mind on this. Modafinil doesn't have the specific problem of caffeine, but it probably has completely different problems. The goal of polyphasic sleep is to use sleep more efficiently. This appears to by skipping early stages of sleep and only getting REM sleep. Quick transition to REM is a symptom of narcolepsy. Narcoleptics have low levels of orexin and orexin antagonists induce REM. Thus it may be that the procedure reduces orexin levels. But modafinil raises orexin, so it may interfere with the main mechanism. Anyhow, there's a testable hypothesis: polyphasic sleep reduces orexin.

Would this be advisable/possible for someone who does moderate to heavy physical activity (lifting heavy at the gym, dancing) almost daily?

3wedrifid8yNot advisable. You will recover more slowly and be more at risk of overtraining.
2Puredoxyk7yBased on my experience and that of people I've spoken to, I would say dancing is fine, lifting is not. I can swim almost infinitely, do martial arts, and even climb moderately with no effect on my sleep; but those things are not tearing muscle generally. Things like lifting and heavy climbing, which tear down / build muscle -- and insanely calorie-intense things like whole days of skin-diving; I learned that one the hard way -- will necessitate extra sleep, often even after adaptation (just like they do when you're monophasic); but during adaptation, they could really screw you up by preventing you from sticking to the schedule. Stay in shape by keeping your cardio and fitness activities, but cut out the serious limits-pushing training for those few weeks, and plan to take some extra naps or a longer core when you do them afterwards, and polyphasic sleep and athletics seem to get along just fine.

I wanted to get to a point where I get REM sleep during naps and tried doing this during my summer vacation. Didn't have a friend to keep me awake. Stayed up one night and napped through the next day fine. The next night I got into a state where I'd start constantly nodding off when sitting down even with the naps every two hours, and was feeling completely out of it. I couldn't figure out how to keep going. I'd been planning on watching TV, which was now right out as I'd completely stopped being able to focus on what's going on, and playing twichy action ... (read more)

Also, what's the point of the fasting, exactly?

Seconded, I'm curious about this.

2Puredoxyk7yThirded; I'm curious about that too. As to brainless activity, cleaning tasks are my favorite. Stupid, picky, clean the shit out of EVERYTHING cleaning tasks. Polishing the silver. Dusting the ceiling. Organizing the socks. Toothbrush-scrubbing the corners. I was lucky enough to stumble on that my first adaptation, and for every one since, or even when I just miss sleep and know I'm going to have a tired night, I make a big list of picky cleaning tasks and just plow through them while I'm tired. They're physical enough to stay awake for, repetitive/stupid enough to not think during, AND they make you feel really great about how you spent the time afterwards!

I don't know how you did your calculations, but:

4 extra hours of productive work or play time per day, or a whole month per year. (Or a decade over 60 years.)

The last two are inconsistent. An extra twelfth of every year, and an extra sixth of every sixty years?

5pinyaka8yMaybe he didn't get enough sleep.

Maybe I didn't get enough sleep.

1LoganStrohl8yIt should be two months per year, since it's another 1/6 of your life spent awake. Thanks for pointing this out! I've fixed it up top.

Do you expect to reach a point where you don't need an alarm clock?

(People already doing this: do you need an alarm clock? How long have you been doing this for?)

5JGWeissman8yI have been doing this for almost a year. I expect to always need an alarm clock. I will often wake up just before the alarm, but this is not reliable. However, I can wake up to very quiet alarm clocks, as long as I recognize the sound as my alarm, so I can sleep through other noise.
3jimmy8yHave you grown to hate the sound of the alarm?
5JGWeissman8yNo, though I have stopped using Ride of the Valkyries because it was too abrupt. I now use the Overture of 1812, which I will sometimes allow to continue playing as I get out of bed.
3[anonymous]8yI used to use "Time" by Pink Floyd.
0Puredoxyk7yI've developed a hilariously pavlovian response to songs I used for alarms at some point or another -- I can still hear "The Authority Song" by Jimmy Eat World and, if I'm sitting or reclining, feel a physical itch to stand. I only use a very quiet beepy thing anymore, or my phone if that's what I've got, and it usually doesn't even go off before I wake up (I deliberately set alarms a few minutes later than I'll wake up so that I have a chance to get up and pre-emptively shut them off), but for a while using songs was a fun way to play with the ol' brain!
0drethelin8yI had this happen with comfort eagle but after switching to more gentle beginning songs I haven't had it happen. I don't do polyphasic though.
4Puredoxyk7yI've been some kind of polyphasic for a solid decade (more, but with breaks that bring it to about that overall). I use an alarm if my schedule is changing -- i.e. I'm doing a day of Uberman to get more done; or I missed a nap and so am sleeping 4.5h tonight instead of 3 -- but even then I often don't need it. Once I'm on my regular Everyman 3 schedule for a few days straight, no alarms are necessary, including popping right awake at 4am feeling great. I only use alarms for naps anymore if I want to read when I wake up, so that I don't get sucked into my book and waste too much time; I wake up so reliably after 20 minutes that my friends have used me as a timer. I love being made of programmable firmware. ;)
0VincenzoLingley7yDo you remember how long it took until you stopped needing alarms?

Brienne, my roommate and I would like to do this, but our schedule won't match yours because we're time constrained. (We already started taking baseline data, but we're doing the adaptation this weekend as opposed to on Monday.) Should we still fill out your form or will the timing mess things up?

Also, my PMV score gets about ~50ms lower when I use a mouse as opposed to a trackpad. Hopefully this isn't an issue, but people should make sure they're using consistent hardware.

3LoganStrohl8yThat's awesome. PM me your names so I know what my data mean.
1benkuhn8yI just realized that my PMV score also varies not just with mouse usage, but also with which OS I'm running (Windows gave me a ~40ms higher score). None of my data is contaminated, but just thought I'd warn you in case anyone at Leverage is dual booting.
1LoganStrohl8yThanks! Yeah, exactly what button is pressed under what conditions is one of the major bug reports I've collected for future research attempts.
3gwern8yFrom the discussion on Yvain's blog of a recent meta-analysis [] trying to compare simple reaction times from Galton to now, I learned that apparently consumer computers have serious issues with displaying images and recording timings down to the ms precision and that one has to configure a setup specifically for the task to compete accuracy-wise with special-purpose reaction-time gear. John Carmack, in video game / VR contexts, has some interesting writings; here's two: [] []

Steve Pavlina on going back to monophasic-- summary: he found that polyphasic was generally good for him mentally and physically (including that he didn't need to cut back on exercise), but he couldn't find a satisfactory way to live with the required tight scheduling.

This article includes a list of all his posts about polyphasic sleep.

Steve Pavlina found that he retained the ability to fall asleep quickly.

2RichardKennaway8yHe also has this recent post [] on how to fall asleep quickly, which is relevant for people on any sleep schedule.
9NancyLebovitz8yPermanent link for Steve Pavlina on how to fall asleep quickly [] . Here's a method I've found for dealing with my occasional insomnia-- I can slow down my heartbeat. I become aware of my heartbeat (pay attention to inner chest area-- trying it now while I'm sitting at the computer, I found it helped to start by using my fingers to feel my pulse at my throat), observe the rhythm, then imagine a slightly slower rhythm, which my heart then follows. I realize this involves some sub-skills. I don't know how common they are, or how hard they are to acquire for people who don't have them.
0Douglas_Knight7yDid you try Pavlina's method?
0Puredoxyk7yOo, nifty info, thanks! I too seemed to retain my ability to fall asleep quickly, no matter what schedule I'm on. But then again, maybe that's because a) I've trained myself so thoroughly to nap and/or b) if I can't sleep pretty quickly, I just get up, no longer having the patience to lay in bed. ;)

I tried to get onto the Uberman sleep schedule twice while at university, but ended up failing around 5 days in. Your plan of easing into it may significantly improve the probability of success, so I am very excited for you.

One thing to watch out for is the increase intake of food. When doing it the first time, I did not think to connect a 37.5% increase in waking time with an increase in calories. I ended up having bouts of extreme coldness in the middle of the night because of this, which added one more thing to the unpleasantness of being awake.

Be caref... (read more)

I've failed Uberman twice myself. You have pretty much an optimal plan, except for the naptation.

"Cut your naps down to 6 as quickly as you can without it hurting too much".

From my own knowledge, which may or may not be trustworthy, naptation doesn't need to be ended prematurely - the whole point is to get a huge number of naps in a short timeframe in order to learn to get REM in a 24-minute interval (which dreaming is a sign of). Getting a few more will just decrease your REM dep. The way I would do it is, get 12 naps a day until you find your... (read more)

[-][anonymous]8y 10

Another thing that happened when I tried this was that no alarm could phase me. Every alarm I tried, including one that required typing my computer password, I would figure out how to turn it off in my sleep. I'm sure I could have continued escalating into solving np complete problems before it stopped, but I gave up soon afterward. I pretty much woke up exclusively from other being physically waking me. I even answered the phone while asleep once, no idea what I said.

5Puredoxyk7yVery good points. Thought I've written a stupid slew on tricks for this sort of thing, my favorite but-I-can-overcome-any-alarm hack is the one I (in one of many moments of silliness) called the Boomstick method: Deeply ingrain a habit of doing some set of activities immediately upon waking. For instance, for a solid month and/or numerous short naps or pseudo-naps, respond to an alarm by leaping up, doing ten jumping jacks, running to the bathroom, slashing water on your face and then reciting a [something]. Then stay awake for a good period of time, and do wakeful things for the first while. You can, in fact, get your body to read a) the alarm as a signal to start that routine, and b) that routine as a signal to flip all the hormones etc. to "I'm awake" position. It's tough to develop, but overall works like a charm.
0arundelo7yI have a guess [] as to what you recited.

What is the point of fasting?

[This comment is no longer endorsed by its author]Reply

What's the point of fasting exactly?

Going to ditto Khoth and hoping for a separate alert - this seems really tempting for me, but I have a strict enough schedule that I don't think I can afford to experiment on low-confidence possibilities.

(Though the recent FB post about switching to lucid dreaming is probably a bad sign...)

I feel I should point out that mindfulness meditation is not something you want to be doing if you're trying to get to sleep. For many people, myself included, doing it before bed can ruin a night's sleep. You want to do something like autogenics if you're trying to nod off, although if you can't, mindfulness meditation might help mitigate the effects of sleep deprivation.

3Puredoxyk7yI agree with you -- I even tried to develop a variant on MM that would be more conducive to allowing sleep to happen if it could, while still being a useful state to just chill in if it couldn't: [] Maybe I hit that balance, maybe I didn't; but I do think that the point is well made that MM itself isn't necessarily the best thing for this purpose.
0zslastman7yInteresting. I'll try it for a few nights and see how it works.

One thing you might want to take into account: your baseline might be bad. Some of the people taking part might already be chronically sleep deprived or have a sleep disorder of some kind or another already.

2Fadeway8yThose people may have a better chance of succeeding.
2Risto_Saarelma8yPuredoxyk, who originated the Uberman idea, was suffering from various sleep disorders [] when she developed the technique. I always did wonder why more people with serious insomnia don't try polyphasic sleep. It's a lot nicer to lie in bed for 20 minutes not falling asleep than it is to lie in bed for 120 minutes not falling asleep.
-1Tenoke8yHe just coined the name. Because if you are suffering from insomnia the last thing you want to do is reduce your daily sleep to potentially 0 minutes since you can reasonably expect to not be getting any sleep from those naps. Additionally, because many people haven't heard of it and because if you are treated by a doctor for your insomnia you will be unlikely to get your doctor's support for it.
3Risto_Saarelma8yShe. And the idea that this is actually something you can get out and train yourself to do wasn't really out there before, even though stuff like Dymaxion sleep had been documented. Your expectance to get some sleep in will probably go up even with most insomnias once you start building up enough sleep deprivation. And there are middle and terminal insomnias that have you able to fall asleep but unable to stay asleep the whole night.
3Puredoxyk7yThank you for the "she". ;) Also, I agree that simply taking data from willing polyphasers means a higher likelihood of your baseline being abnormal for one reason or another, the presence of already-screwy sleep definitely among them. However, the research has gotta start somewhere, and I'm thrilled that voluntary groups are starting to form -- it's a great step, and LW are just about the perfect people to be on it IMO. <3 (I did "just name it", but before I started writing about it in '00, there was no data other than Dr. Fuller's, which was really only recorded / disseminated in a tiiiiiiny Time Magazine blurb in the 80's; and Dr. Stampi's work on naps generally, which is excellent but limited and doesn't discuss polyphasic sleep as a lifestyle. I didn't know this at the time -- I thought I was just writing my experiences with something that surely other people had written more about elsewhere...which I mention to excuse the rather slipshod quality of my early work. It's why I tried so hard to improve the book and to make my site more comprehensive later...I hope I've made up for some of the early lapses.) Thanks!